

Mixtures of Etltane and Nitrous Oxide. 175 



condensation, was explained bv retardation*, which was 

 annihilated in my experiments for the first time by means of 

 a small iron stirring-rod within the tubes, which was moved 

 by an electromagnet outside the tubes. 



5. If P lies on the other side of R there are again two 

 definite critical temperatures, Tp and T R , for every mixture ; 

 but the process of condensation between those two will be 

 different from what it was in the first case. By compression 

 a new phase appears of a larger volume f this time, i. e. a 

 vapour-phase, which begins by increasing, then reaches 

 a maximum, and then again diminishes and disappears. 

 This was called retrograde condensation of the second hind 

 (r.c. II.) f. Near the plaitpoint-temperature T P the liquid 

 will be very flat and indistinct, and more so at the beginning 

 than near the end of the condensation. Below T P the con- 

 densation is normal. 



6. This second case has never been realized till now. I 

 have been considerino- whether it might be, and the choice of 

 the subject of this paper is the outcome of my reasoning. 

 In explaining this 1 shall use a different graphical represen- 

 tation, namely the p-t figure. In a recent paper § I have 

 deduced the general features of the p-t curves in the problem 

 before us from van der Waals's theory. The figure obtained 

 consists of (I) the two vapour-pressure curves of the compo- 

 nent substances ; (2) the border-curves || for the mixtures 

 [x = constant) having the form of loops ; (3) the plaitpoint- 

 curve, being the envelope of the border-curves and connecting 

 the critical points of the two substances. The point of con- 

 tact between a border-curve and the plaitpoint-curve corre- 

 sponds to the plaitpoint P of van der Waals's surface ; the 

 critical point of contact R is the point in which the border- 

 curve has a tangent parallel to the jo-axis. When going along 

 the border-curve, starting from its lower branch, we may 

 either first reach R or P. These two cases correspond to the 



* Kuenen, Archives Neerl xxvi. pp. 376-377, 380-382; Commim. 

 No. 4; Zeitschr. phys. Chemie, xi. p. 4/. 



t In order to make the volumes in our figures decide about the situa- 

 tion of the phases above each other, we should take the volume of the 

 unit of weight as the unit of volume. In doing so a phase of larger 

 volume is a lighter phase at the same time, and will therefore appear in 

 the upper part of the tube. 



\ Kuenen, Archives Neerl. xxvi. pp. 388-389 ; Zeitschr. phys. Chemie, 

 xi. p. 45. 



§ Kuenen, Kon. Akad. Amsterdam, 29 Sept. 1894, pp. 90-99 ; ' Com- 

 munications,' &c. No. 13. 



|| A border-curve for a mixture represents the conditions (p, t) in 

 which it may coexist with a second phase. 



